/!\ The raw file t01-02 is not filtered and contains a lot of wrong
    signal strength (127 dBm).
    The other files have been filtered:
    - all the autocalibration requests containing at least one
      occurrence of 127 dBm have been deleted;
    - positioning requests have been deleted if at the end of a phase or
      if they contained a majority of 127 dBm packets, otherwise only
      the wrong packets have been removed from the request.

t01 is the path towards CP5, and t02 is the return, from CP5 to the
emergency exit. Both tests were done in a row and generated only one
big aggregation file.

t01 started approximately at 13:07, and t02 finished approximately at
13:30, but some time passed between the two tests:

Timestamp of the last request of t01: 1368650837.872503991
Timestamp of the first request of t02: 1368651254.630584995
Difference: 416 seconds = about 6 minutes 56 seconds

t01 finished approximately at 13:15, and t02 started around 13:22.

Each test itself lasted about 7 minutes 30 seconds. See the dedicated
report files for precise timing.